Adaptive play via estimation in uncertain nonzero-sum orbital pursuit evasion games

2017, American Institute of Aeronautics and Astronautics Inc, AIAA. All rights reserved. Much study has been given to the topic of pursuit-evasion games, which are of interest in diverse fields including engineering, artificial intelligence, and economics. Attention has focused largely on complete information games with perfect information. In this paper, an incomplete, imperfect information game is considered. Two spacecraft are engaged in a game about a circular orbit, utilizing the Hill-Clohessy-Wiltshire equations of relative motion. Each agent measures the system state at a fixed rate with corrupting white noise, and has partial knowledge of the opponents linear-quadratic objective function, subject to parametric uncertainty. Sequential estimation is employed to obtain knowledge of the opponents objective function weights, allowing each agent to transform the problem into a one-sided optimization problem. The adaptive strategies are presented for a pursuer playing a static evader, an evader playing a static pursuer, and for simultaneously adapting agents. Simulation results are shown.

Adaptive play via estimation in uncertain nonzero-sum orbital pursuit evasion games Conference Paper